Pull ya DAX up!

[CanOz]

Sam, really good hear. You might want to consider these guys for funding later on www.mescapitalgroup.com

I'm working my way through a trial with them now, if i pass they'll fund me, so they say. Although at the moment i have to trade CME products, they say i can trade my markets later. They're very hands off, unlike TopStepTrader

I'm only trading the 6A and the 6J, for now.

Cheers,


CanOz

 

[ThingyMajiggy]

Sam, really good hear. You might want to consider these guys for funding later on www.mescapitalgroup.com

I'm working my way through a trial with them now, if i pass they'll fund me, so they say. Although at the moment i have to trade CME products, they say i can trade my markets later. They're very hands off, unlike TopStepTrader

I'm only trading the 6A and the 6J, for now.

Cheers,


CanOz

Interesting, never heard of them. Will check it out, thanks for the heads up

 

[ThingyMajiggy]

I have decided I will post more in here again. I haven't traded for a while and when I did I was trading with CFDs, but the goal is futures trading. I haven't traded for a few months now as I have been away on holiday, have broken my wrist(which halted the saving/trading) so it's taken a bit of a back seat, I was also putting some time into learning programming languages(Python, Javascript, C++ etc).

I have decided I am going to try and be as honest as I can with myself and my trading and try and find my way properly, not copy someone else or try to incorporate someone else's system, but do what feels good and right for me.

To me, I have always thought to myself that order flow does indeed make the most sense, I tend to think if I can see every order going through the market, surely there are patterns and games there that I can see if I just spend the time watching to try and pick up on these things. Some other soft spots I have towards what makes sense, I tend to like the idea of deviation trading and mean-reversion style trading.

I once made an indicator on CQG when I was using that platform that would draw lines every X ticks apart above(of the users choosing) at the open that would be the daily standard deviation, hourly deviation, weekly deviation etc.(I would have multiple plots of this indicator) that I would work out in excel, so get the daily data of the instrument, get the daily ranges, avg range, std dev and use that figure as the distance apart to automatically plot the lines. My thinking, probably wrong, is that we know things usually don't(or rare) have days that are outliers, like 3 or 4 standard deviations, so I thought I'd represent this as lines from each day's open and get a feel for how many standard deviations we have/haven't gone today. I have since adapted this to Ninjatrader just to see how it looks and how well the lines do, like so, still very much a work in progress and just seeing how they go, not really making a strategy or method to trade with, just curiosity at this point, this is today's session so far, so these were drawn on the open of each session this week, so 2 days, they're just spanning over into each others session(something I'm working on):



Other than that I've been playing around with watching the sessions back at 8x speed using BookMap and nothing else, just to see what I can see, have tweaked a few settings and put in trailing stops which you can see here which is a 10 tick stop initially then trails it as the trade goes in favour, although the result is good(+2762 eur up the top) it's kind of hard to explain what exactly I was doing, which feels weird to be honest, it feels I should be more systematic in my approach, but here I am just trading by feel and I guess getting in on pullbacks or where liquidity was, but certainly no rules like when this line crosses that line type stuff at all, another thing that bugs me is that doing replay stuff is a bit skewed, as you already know in the back of your head how the day is going to pan out, so its easy to be trading for how the day pans out rather than actually looking for good entries/legit things to get into trades for, which is something I need to work on. But this is what that looks like anyway, little blue marks are entry, pink line is trailing stop automatically moving up for me:



Whether any of this will morph into something worthwhile is yet to be seen, as I mentioned before this is all just my own way of trying to find something that works for me, consistently, and to learn and watch how my market moves.

I also want to learn to be much much better at managing trades, have been trying to find stuff about taking completely random entries and still making it work via risk management. Seems to be that the better you are at managing trades and risk, the higher the chance that you will be a profitable trader, choosing what to focus on(entries vs exits) etc. I need help in this area, all areas of the above actually but hopefully I can figure some of it out that suits me.

I have also been jotting down things on a notepad that I'm trying, keeping track of at least what I'm trying so I can then focus more on things that show a little bit of promise, rather than swapping and changing between things all the time.

Excuse the enormous post

 

[CanOz]

Bookmap and auction Vista are the future here and now in my view, it's like shining a torch on the thief in the room. It's very effective for me on fast markets and I wonder what my stats would have looked like if I started using it sooner! Using on the HSI now and it works OK with IBs data.

 

[ThingyMajiggy]

Bookmap and auction Vista are the future here and now in my view, it's like shining a torch on the thief in the room. It's very effective for me on fast markets and I wonder what my stats would have looked like if I started using it sooner! Using on the HSI now and it works OK with IBs data.

Yeah I've had BookMap for a while, pretty impressive stuff, very often updated too, has come a long way since it first started with great features being added. It's definitely an interesting viewpoint to look at the order flow, not sure if it's a bit of a looks more helpful than it actually is thing yet though. Time will tell But it has fantastic replay abilties and a fantastic way of viewing the order flow that's for sure.

 

[ThingyMajiggy]

Alright, so I've done more on my SamLines got it looking more how I want it now, much better. So here is what they look like, so you have the blue lines above, red lines below, all the lines are automatically drawn on the open so I'm not putting them in after the fact or anything. So I've zoomed right out looking at the last 21 days or so I can get an idea of what might be happening, and can see that around 90% of the days move one standard deviation away from its opening price, I've looked back and found that days when it doesn't touch either one stdev up or one stdev down, it seems to do it two days in a row, which is weird, but here is the zoomed out look:



You're probably wondering what the green and pink lines are, so I've also calculated the standard deviation of the ranges based on hourly data as well, which is what the green and pink version is, so a more short term zoomed in effect for minute data etc. So I've gone through a just quickly picked a few days that give examples.

First one is last night's session:



Next one is a day where it didn't reach one daily standard deviation up or down, so just bounced around between the hourly stdev:



And the last one is a day where it hit both, one down then bang straight up to one up, there was also a day where it went almost 2 up then smashed down to 3 down but obviously that's rare, that was like a 600 point move, this one is just touching both first levels(first red and first blue in same session):



So it's interesting to look at if nothing else I know there's a fair few lines going on and it might be a bit of having that many lines so of course it's going to look like its bouncing off some of them I guess, but still they seem to point out areas that something at least happens, or it pauses, or turns completely. I would like to test somehow to see percentage chance of hitting the 2nd levels, or if there might be any patterns as far as if we hit level 2 one day, is it likely to hit the opposite first level the next day or something, you get my drift, but not quite sure how to go about that just yet.

Have got another thing to post as well, just a little data export indicator that I whacked up quick and nasty, but will do another post otherwise it will be too many images/attachments I think.

 

[ThingyMajiggy]

Another thing that was bugging me was trying to get hourly data, NT's export feature is either daily, minute or tick which wasn't what I was after. So I just made my own very simple nasty "indicator" that does the trick. So I'll attach that here in case anyone else might like to use it or extracting data. So it works by just importing the ninjascript, displaying the output window in NT by going to Tools >Output window like so:



So that's just a normal 60 minute chart, nothing else on it, it doesn't matter if you did have indicators all over it or whatever, would still work. Next to it is NTs output window. So then you just add the indicator on the chart and it will fill the output window with Time, Open, High, Low, Close data, like so:



Then you can just right click in the output window, choose Save As, then open it up in Excel on which you will get this window where you choose delimited:



Then you select "comma" as the delimiter, which will separate the data into columns, then next and finish and tadaaaa like so:



Then it looks like this in excel:



So you can export any data you like, 1min, 5min, hourly etc. Just load up how much history you want in your chart settings and run this "indicator" with the output window open so you can then save it and open in Excel

 

[ThingyMajiggy]

Annnd now to attach the actual indicator Run out of attachments again I think. It's basically just a Print output statement with the TOHLC details, nothing fancy, but it does the job easily.

File > Utilities > Import Ninjascript for those that don't know how.

 

[ThingyMajiggy]

Off to a good start, thingylines being nice to me so far + gap close, plus I'm kind of expecting today to be a little bit of a quieter retrace day after two fairly strong days but we shall see, now it's time to panic and give it all back

 

[Modest]

Nice trade!

 

[CanOz]

Great work with the std dev bands Sam. This has the makings of a statistical edge if you can find out what the distribution of results would look like over a meaningful sample.

Staying tuned!

 

[ThingyMajiggy]

Great work with the std dev bands Sam. This has the makings of a statistical edge if you can find out what the distribution of results would look like over a meaningful sample.

Staying tuned!

Distribution of what results? If it lies according to the std dev rule of thumb? Like 95% of the time between 2 standard deviations etc. or if everyday touches a first layer? It seems to be giving me good expectations anyway, helping in that department, helping me plan how the day might play out

Thanks Mod and Can

Did alright last night, +120 ticks. Night played out pretty much spot on with my expectations which was nice, an off day after two strong days up to touch the first lower band, bit of a retrace day. Probably could have done much better, but got that first trade from the high of day which I posted last night, then a couple more later from the lows but would have been nicer to get all the down move, but got a nice chunk to the gap close which was a safe call, got to have the gonads to hold through on my expectations though, but still happy with that. I'm off for the weekend now so won't get to watch tonight, have a good one all

 

[CanOz]

Distribution of what results?

Well if you could actually determine (i.e. through testing or analyzing the stats) what % of the time the market trades at, over and under the daily range then you have the making of a statistical edge. If you knew for example that only 10% of the time does the market trade beyond its daily average range by x%, then you could apply that to another play, like a mean reversion pattern at a level etc...

As an example, if the ES opens within x% of its globex high or low, there is a 90+ % chance that reference will be taken out. That's a statistical edge. I thought you might be going that way with the bands and i was looking forward to the results. If not, that's fine....carry on

 

[ThingyMajiggy]

Yeah that's cool, just was a bit confused with the way you worded it. That's what I thought you meant. That's definitely something I hope to find out, just not sure of how. I know that with standard deviation in a normal distribution its something like 95% that it stays at 2 standard deviations or below so might be good to take turning points at beyond the 2nd level etc. Could come in handy

 

[Joules MM1]

Yeah that's cool, just was a bit confused with the way you worded it. That's what I thought you meant. That's definitely something I hope to find out, just not sure of how. I know that with standard deviation in a normal distribution its something like 95% that it stays at 2 standard deviations or below so might be good to take turning points at beyond the 2nd level etc. Could come in handy

are sigmas worth a study for this idea ?

 

[rvm]

are sigmas worth a study for this idea ?

Yes, it's useful though I haven't traded CFDs often with this idea.

However, I have successfully tried this with certain bluechips and mid-cap equities that get clobbered/elevated by macro(instead of news). Found this handy during the 'boring' season(between May and October) when stock prices trade ranges in the absence of news. It's kind of a low-risk way to trade high probability edge cases.

I am currently working on a way to automate the process and incorporate other signals, but my programming skill is poor and I haven't had much spare time. I've requested some help - see how that turns out.

 

[Joules MM1]

Yes, it's useful though I haven't traded CFDs often with this idea.

However, I have successfully tried this with certain bluechips and mid-cap equities that get clobbered/elevated by macro(instead of news). Found this handy during the 'boring' season(between May and October) when stock prices trade ranges in the absence of news. It's kind of a low-risk way to trade high probability edge cases.

I am currently working on a way to automate the process and incorporate other signals, but my programming skill is poor and I haven't had much spare time. I've requested some help - see how that turns out.

Fari Hamzei (constant top 5 US market timers according to timerdigest.com) is now making calls on $DAX
he's a big user of sigmas ...take some time to check out his work....at least watch a cupla vids of his if you want a diff perspective

https://twitter.com/HamzeiAnalytics
http://www.hamzeianalytics.com/Educational_Webinars.asp

 

[Gringotts Bank]

Guys, range bands based on historical data will be very curve fitted. Remember to do a walk forward test.

 

[ThingyMajiggy]

Guys, range bands based on historical data will be very curve fitted. Remember to do a walk forward test.

Isn't everything curve fitted though? All just based on recent price history no matter what indicator/tool it is?

And joules sigma is std dev? Same thing?

 

[Gringotts Bank]

Isn't everything curve fitted though? All just based on recent price history no matter what indicator/tool it is?

And joules sigma is std dev? Same thing?

Yeh everything is fitted, but not everything works on unseen data.

I had to look up sigma. Six sigma came up a bit on google - basically a big move (6 std devs)and what happens next. http://www.marketcalls.in/trading-l...-about-six-sigma-events-in-stock-markets.html